• The Journal of the Acoustical Society of Korea
• /
• v.26 no.6
• /
• pp.259-268
• /
• 2007

This paper contributes to an improvement of the statistical bandwidth extension(BWE) system based on Hidden Markov Model(HMM). First, the existing HMM training method for BWE, which is suggested originally by Jax, is analyzed in comparison with the general Baum-Welch training method. Next, based on this analysis, a new HMM-based BWE method is suggested which adopts the Baum-Welch re-estimation algorithm instead of the Jax's to train HMM model. Conclusionally speaking, the Baum-Welch re-estimation algorithm is a generalized form of the Jax's training method. It is flexible and adaptive in modeling the statistical characteristic of training data. Therefore, it generates a better model to the training data, which results in an enhanced BWE system. According to experimental results, the new method performs much better than the Jax's BWE systemin all cases. Under the given test conditions, the RMS log spectral distortion(LSD) scores were improved ranged from 0.31dB to 0.8dB, and 0.52dB in average.

• The Journal of the Acoustical Society of Korea
• /
• v.27 no.7
• /
• pp.350-356
• /
• 2008

In this paper, we propose shading functions to the appropriate focused beamforming for near-field target estimation. This near field shading functions are based on Chebychev and Manning windows. In order to obtain the optimum sensor weighting values with the help of the proposed shading technique, we assume that the sensor positions associated to the non-uniformly distributed array are precisely known. We calculate a series of sensor weighting values from the FFT operation of given shading functions in time domain. By applying the shading weights on the sensor array, we can see that the level of sidelobe becomes diminished and the performance of estimating range and azimuth gets improved. In addition, we propose a non-uniform structure in terms of frequency bands, which may minimize the attenuation of incoming signals.

• Journal of the Korea Computer Graphics Society
• /
• v.14 no.3
• /
• pp.31-46
• /
• 2008

The depth of field is the range that the objects inside of this range treated to be focused. Objects that are placed out of this range are out of focus and become blurred. In computer graphics, generating depth of field effects gives a great reality to rendered images. The previous researches on the depth of field in computer graphics can be divided into two major categories. One of them is the distributed ray tracing that samples the lens area against each pixel. It is possible to obtain precise results without noise if enough number of samples are taken. However, to make a good result, a great number of samples are needed, resulting in an enormous timing requirement. The other approach is the method that approximates depth of field effect by post-processing an image and its depth values computed using a pin-hole camera. Though the second technique is not that physically correct like distributed ray tracing, many approaches which using this idea have been introduced because it is much faster than the first approach. But the post-processing have some limitations because of the lack of ray information. In this paper, we first present an improvement technique that corrects the previous post-processing methods and then propose another one that accelerates the distributed ray tracing by using a radiance caching method.

• Journal of Internet Computing and Services
• /
• v.16 no.4
• /
• pp.39-50
• /
• 2015

In information retrieval process through search engine, some users want to retrieve several documents that are corresponding with specific time period situation. For example, if user wants to search a document that contains the situation before 'Japanese invasions of Korea era', he may use the keyword 'Japanese invasions of Korea' by using searching query. Then, search engine gives all of documents about 'Japanese invasions of Korea' disregarding time period in order. It makes user to do an additional work. In addition, a large percentage of cases which is related to historical documents have different time period between generation date of a document and record time of contents. If time period in document contents can be extracted, it may facilitate effective information for retrieval and various applications. Consequently, we pursue a research extracting time period of Joseon era's historical documents by using historic literature for Joseon era in order to deduct the time period corresponding with document content in this paper. We define historical objects based on historic literature that was collected from web and confirm a possibility of extracting time period of web document by machine learning techniques. In addition to the machine learning techniques, we propose and apply the similarity filtering based on the comparison between the historical objects. Finally, we'll evaluate the result of temporal indexing accuracy and improvement.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.20 no.1
• /
• pp.65-73
• /
• 2007

In the inverse perturbation method, enormous computational resource was required to obtain reliable results, because all unspecified DOFs were considered as unknown variables. Thus, in the present study, a reduced system method is used to condense the unspecified DOFs by using the specified DOFs, and to improve the computational efficiency as well as the solution accuracy. In most of the conventional reduction methods, transformation errors occur in the transformation matrix between the unspecified DOFs and the specified DOFs. Thus it is hard to obtain reliable and accurate solution of inverse perturbation problems by reduction methods due to the error in the transformation matrix. This numerical trouble is resolved in the present study by adopting iterative improved reduced system(IIRS) as well as by updating the transformation matrix at every step. In this reduction method, system accuracy is related to the selection of the primary DOFs and Iteration time. And both are dependent to each other So, the two level condensation method (TLCS) is selected as Selection method of primary DOFs for increasing accuracy and reducing iteration time. Finally, numerical verification results of the present iterative inverse perturbation method (IIPM) are presented.

• Korean Journal of Remote Sensing
• /
• v.35 no.6_3
• /
• pp.1209-1219
• /
• 2019

For consistent vegetation monitoring, it is necessary to generate time-series vegetation index datasets at fine temporal and spatial scales by fusing the complementary characteristics between temporal and spatial scales of multiple satellite data. In this study, we quantitatively and qualitatively analyzed the prediction accuracy of time-series change information extracted from spatio-temporal fusion models of multiple satellite data for vegetation monitoring. As for the spatio-temporal fusion models, we applied two models that have been widely employed to vegetation monitoring, including a Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) and an Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model (ESTARFM). To quantitatively evaluate the prediction accuracy, we first generated simulated data sets from MODIS data with fine temporal scales and then used them as inputs for the spatio-temporal fusion models. We observed from the comparative experiment that ESTARFM showed better prediction performance than STARFM, but the prediction performance for the two models became degraded as the difference between the prediction date and the simultaneous acquisition date of the input data increased. This result indicates that multiple data acquired close to the prediction date should be used to improve the prediction accuracy. When considering the limited availability of optical images, it is necessary to develop an advanced spatio-temporal model that can reflect the suggestions of this study for vegetation monitoring.

• Journal of KIISE:Information Networking
• /
• v.36 no.6
• /
• pp.514-527
• /
• 2009

Wireless sensor networks (WSNs) can be used in various applications for military or environmental purpose. Recently, there are lots of on-going researches for detecting and tracking the spread of continuous objects or phenomena such as poisonous gas, wildfires, earthquakes, and so on. Some previous work has proposed techniques to detect edge nodes of such a continuous object based on the information of all the 1-hop neighbor nodes. In those techniques, however, a number of nodes are redundantly selected as edge nodes, and thus, the boundary of the continuous object cannot be presented accurately. In this paper, we propose a new edge detection method in which edge nodes of the continuous object are detected based on the information of the neighbor nodes obtained via the Localized Delaunay Triangulation so that a minimum number of nodes are selected as edge nodes. We also define the sensor behavior rule for tracking continuous objects energy-efficiently. Our simulation results show that the proposed edge detection method provides enhanced performance compared with previous 1-hop neighbor node based methods. On the average, the accuracy is improved by 29.95% while the number of edge nodes, the amount of communication messages and energy consumption are reduced by 54.43%, 79.36% and 72.34%, respectively. Moreover, the number of edge nodes decreases by 48.38% on the average in our field test with MICAz motes.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.48 no.4
• /
• pp.6-14
• /
• 2011

The fast growing of the number of users requires the development of reliable communication systems able to provide higher data rates. In order to meet those requirements, techniques such as Multiple Input Multiple Out (MIMO) and Orthogonal Frequency Division multiplexing (OFDM) have been developed in the recent years. In order to combine the benefits of both techniques, the research activity is currently focused on MIMO-OFDM systems. In addition, for a fast wireless channel environment, the data rate and reliability can be optimized by setting the modulation and coding adaptively according to the channel conditions; and using sub-carrier frequency, and power allocation techniques. Depending on how accurate the feedback-based system obtain the channel state information (CSI) and feed it back to the transmitter without delay, the overall system performance would be poor or optimal. In this paper, we propose a Signal to Noise Ratio (SNR) estimation algorithm where the preamble is known for both sides of the transciever. Through simulations made over several channel environments, we prove that our proposed SNR estimation algorithm is more accurate compared with the traditional SNR estimation. Also, We applied AMC on several channel environments using the parameters of IEEE 802.11n, and compared the Throughput performance when using each of the different SNR Estimation Algorithms. The results obtained in the simulation confirm that the proposed algorithm produces the highest Throughput performance.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.10
• /
• pp.15-23
• /
• 2016

To support the telematics and infotainment services, vehicle-to-everything (V2X) communication requires a robust and reliable network. To do this, the 3rd Generation Partnership Project (3GPP) has recently developed V2X communication. For reliable communication, accurate channel estimation should be done. However, because vehicle speed is very fast, radio channel is rapidly changed with time. Therefore, it is difficult to accurately estimate the channel. In this paper, we propose the new linear minimum mean square error (LMMSE) channel interpolation scheme based on the Long Term Evolution (LTE) sidelink system in vehicle-to-vehicle (V2V) environments. In our proposed reduced decision error (RDE) channel estimation scheme, LMMSE channel estimation is applied in the pilot symbol, and then in the data symbol, smoothing and LMMSE channel interpolation scheme is applied. After that, time and frequency domain averaging are applied to obtain the whole channel frequency response. In addition, the LMMSE equalizer of the receiver side can reduce the error propagation due to the decision error. Therefore, it is possible to detect the reliable data. Analysis and simulation results demonstrate that the proposed scheme outperforms currently conventional schemes in normalized mean square error (NMSE) and bit error rate (BER).

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.7A
• /
• pp.586-599
• /
• 2012

This paper proposes a fast face detection and tracking method which uses depth images as well as RGB images. It consists of the face detection procedure and the face tracking procedure. The face detection method basically uses an existing method, Adaboost, but it reduces the size of the search area by using the depth image. The proposed face tracking method uses a template matching technique and incorporates an early-termination scheme to reduce the execution time further. The results from implementing and experimenting the proposed methods showed that the proposed face detection method takes only about 39% of the execution time of the existing method. The proposed tracking method takes only 2.48ms per frame with $640{\times}480$ resolution. For the exactness, the proposed detection method showed a little lower in detection ratio but in the error ratio, which is for the cases when a detected one as a face is not really a face, the proposed method showed only about 38% of that of the previous method. The proposed face tracking method turned out to have a trade-off relationship between the execution time and the exactness. In all the cases except a special one, the tracking error ratio is as low as about 1%. Therefore, we expect the proposed face detection and tracking methods can be used individually or in combined for many applications that need fast execution and exact detection or tracking.